ctmmweb

ctmmweb implements continuous-time speed and distance (CTSD) estimation to produce scale-insensitive estimates of speed and distance traveled from animal tracking data.

Key Features:

• Continuous-time speed and distance (CTSD): Implements CTSD estimation to derive speed and distance metrics in continuous time rather than by summing straight-line displacements (SLDs).
• Separation of processes: Separates the underlying continuous-time movement process from the discrete-time sampling process to reduce sensitivity to sampling schedule.
• Measurement error calibration: Estimates device error parameters to calibrate and account for measurement error in tracking data.
• Model selection: Applies model selection techniques to identify the optimal continuous-time movement model for the data.
• Simulation-based trajectory sampling: Uses simulation to sample trajectories conditional on the observed data distribution.
• Estimation with uncertainty: Extracts mean speed estimates accompanied by confidence intervals for formal inference.
• Robustness to sampling frequency and tortuosity: Produces estimates that are less sensitive to sampling frequency and movement tortuosity than SLD methods.
• Empirical and synthetic validation: Validated using simulations with synthetic data and applied to empirical GPS tracking data.

Scientific Applications:

• Speed and distance estimation: Quantifies animal movement speed and distance traveled from tracking datasets.
• Uncertainty quantification: Provides confidence intervals for speed and distance estimates to enable formal statistical inference.
• Bias correction: Corrects biases arising from sampling frequency, tortuosity, and measurement error that affect SLD-based estimates.
• Model-based movement analysis: Facilitates selection and application of continuous-time movement models to tracking data.
• GPS data analysis: Applies CTSD methods to empirical GPS tracking datasets and to synthetic simulation studies.

Methodology:

Calibrate measurement error by estimating device error parameters; perform model selection to identify an optimal continuous-time movement model; use simulation-based sampling of trajectories conditional on observed data to extract mean speed estimates and associated confidence intervals while separating the continuous-time movement process from the discrete-time sampling process.